summaryrefslogtreecommitdiff
path: root/gnome-settings-daemon
Commit message (Collapse)AuthorAgeFilesLines
* main: Use sh not bash for the localeexec helperBastien Nocera2013-06-131-2/+3
| | | | | | Replace the bash-only substitution with an sh-compatible version. https://bugzilla.gnome.org/show_bug.cgi?id=701322
* main: Fix incorrect quote removalBastien Nocera2013-06-131-1/+1
| | | | | | We were trying to remove a comma instead of a single-quote. https://bugzilla.gnome.org/show_bug.cgi?id=701322
* Avoid calling setenv after starting threadsMatthias Clasen2013-06-103-33/+28
| | | | | | | | | | | In particular, we can't get the locale value out of GSettings in order to set LC_PAPER etc, since calling into GSettings initializes the dconf backend which in turn uses gdbus, which starts a worker thread. As a simple workaround, set up the locale environment in a small wrapper script that then exec's the g-s-d binary. https://bugzilla.gnome.org/show_bug.cgi?id=701322
* main: fix buildWilliam Jon McCann2013-02-151-39/+41
|
* main: Set our own IBus envvars as wellBastien Nocera2013-02-151-1/+9
| | | | | | Just as we set gnome-session's. https://bugzilla.gnome.org/show_bug.cgi?id=693381
* main: Set locale envvars on startupBastien Nocera2013-02-151-0/+23
| | | | | | | | When starting up, we tell gnome-session to set the environment variables for new starting programs, but we don't set the environment for our own children. Do this on start up, as soon as possible. https://bugzilla.gnome.org/show_bug.cgi?id=693381
* daemon: Also apply LC_PAPERBastien Nocera2013-01-261-0/+1
| | | | | | User-specified format should also apply to the paper format. https://bugzilla.gnome.org/show_bug.cgi?id=690750
* Fix PluginName handling in new macroMatthias Clasen2013-01-021-1/+1
| | | | | | When constructing a string literal, one needs to use stringification, not tokenization. This was causing PluginName## to appear in debug output.
* main: Use new defines for D-Bus namesBastien Nocera2012-12-192-8/+5
|
* main: Export D-Bus name and interface to pluginsBastien Nocera2012-12-191-0/+4
|
* main: Start plugins as early as possibleBastien Nocera2012-12-171-2/+1
| | | | | | Even before we've registered with gnome-session, to avoid the window manager starting before some of the necessary early plugins such as the cursor plugin.
* plugins: Simplify plugin creation with a huge macroBastien Nocera2012-12-131-5/+75
| | | | Shaves about 2.7k lines of code.
* main: exit gracefully if the session name is already takenPeter Hutterer2012-12-121-1/+5
| | | | https://bugzilla.gnome.org/show_bug.cgi?id=690078
* main: Add a way to replace the daemonJasper St. Pierre2012-11-301-1/+11
| | | | | | | This is intended for quickly and easily testing g-s-d changes without having gnome-session restart it. https://bugzilla.gnome.org/show_bug.cgi?id=689402
* main: Fix early morning codeBastien Nocera2012-11-301-1/+2
|
* main: Don't set IBus envvars if keyboard plugin is disabledBastien Nocera2012-11-301-1/+14
| | | | https://bugzilla.gnome.org/show_bug.cgi?id=689070
* main: Add settings key to whitelist pluginsBastien Nocera2012-11-141-0/+21
| | | | | | | | | | | Rather than having gdm load all the new plugins by default, especially when gnome-settings-daemon gets updated and gdm doesn't, keep a list of whitelisted plugins. Only those can be enabled, others will be unknown and ignored. By default, all plugins are whitelisted. https://bugzilla.gnome.org/show_bug.cgi?id=662220
* main: remove a systemd header that was missedRyan Lortie2012-11-121-1/+0
| | | | | | | Remove an #include that was missed by Colin's patch to remove the systemd dependency. https://bugzilla.gnome.org/show_bug.cgi?id=686650
* main: Remove libsystemd-login specific codeColin Walters2012-11-122-265/+0
| | | | | | | | | | | | After porting both users of the helper code (the power and color plugins) to use gnome-session's SessionIsActive property, remove the libsystemd-login dependent code. As originally intended, gnome-settings-daemon requires systemd for the power plugin to work at run-time, but does not hard depend on libsystemd-login itself at compile-time. https://bugzilla.gnome.org/show_bug.cgi?id=686650
* main: Always set the IBus envvarsBastien Nocera2012-11-091-16/+3
| | | | Only if IBus is available though. #fallback
* Remove calls to g_type_init(), bump GLib required versionColin Walters2012-10-291-2/+0
| | | | | | See http://git.gnome.org/browse/glib/commit/?id=7c42ab23b55c43ab96d0ac2124b550bf1f49c1ec https://bugzilla.gnome.org/show_bug.cgi?id=687072
* main: Only start the manager after registering with gnome sessionRui Matos2012-10-261-32/+28
| | | | | | | | This ensures that any plugin idles will only run after RegisterClient but also that we only call RegisterClient after Setenv since otherwise Setenv isn't effective. https://bugzilla.gnome.org/show_bug.cgi?id=686814
* Revert "main: Ensure session registration happens before other idles"Rui Matos2012-10-261-44/+0
| | | | | | This reverts commit dcb54d8a1c4ecacf6fbe17bdffefdca3b1e6d2d5. https://bugzilla.gnome.org/show_bug.cgi?id=686814
* main: Link to systemd againColin Walters2012-10-221-0/+1
| | | | Commit e76a752c7d3901b3abbc1733b824c80a3dddec16 broke it.
* main: Make GnomeSettingsSession a singletonMatthias Clasen2012-10-221-3/+10
| | | | | | | | There is no need at all to do the listening for systemd or ConsoleKit signals more than once in the gnome-settings-daemon process. https://bugzilla.gnome.org/show_bug.cgi?id=686554
* main: Those flags are now included in the daemon's flagsBastien Nocera2012-10-221-2/+0
|
* main: Require logind for session trackingBastien Nocera2012-10-221-183/+4
| | | | | | | | Instead of offering both ConsoleKit and logind support. We already require logind being available for inhibition and suspend/hibernation in the power plugin, so requiring logind for session tracking doesn't add any new dependencies.
* main: Ensure session registration happens before other idlesMatthias Clasen2012-10-221-0/+44
| | | | | | | | | | | | | | We're deferring some slower operations to idles, in plugins, to have it happen after session registration and not slow down login - but I've observed that the ordering is unreliable. Sometimes we end up doing the RegisterClient call after the initial set of idles. Ensure that the session registration happens before other idles by making the idle higher priority than the plugins. https://bugzilla.gnome.org/show_bug.cgi?id=686579
* main: Use new SessionManager singletonMatthias Clasen2012-10-221-82/+25
| | | | https://bugzilla.gnome.org/show_bug.cgi?id=686556
* main: Provide a singleton SessionManager proxy objectMatthias Clasen2012-10-222-0/+33
| | | | | | | | Many plugins were creating these on their own, with ugly async code. It seems much more straightforward and efficient to get a proxy once, and keep it around for everybody to use. https://bugzilla.gnome.org/show_bug.cgi?id=686556
* main: Don't set Input Methods envvar without IBusBastien Nocera2012-10-051-1/+14
| | | | | | | | | | | If IBus isn't installed, don't set the QT_IM_MODULE or XMODIFIERS envvars, as that breaks other Input Methods. As we advice people who don't want to use IBus to just uninstall ibus itself, we should handle being built with IBus support but run-time support not being there. https://bugzilla.gnome.org/show_bug.cgi?id=685514
* main: Set session env vars to make non gtk+ apps work with IBusRui Matos2012-09-031-15/+77
| | | | | | | | | If we are compiled with IBus support and the current session isn't fallback then we always set the QT_IM_MODULE and XMODIFIERS environment variables so that Qt and XIM supporting applications work as best as possible. https://bugzilla.gnome.org/show_bug.cgi?id=680313
* plugins: add class_finalize to REGISTER macroRay Strode2012-08-211-1/+6
| | | | | | | | We already have a convenience macro for plugins to use to define their types. This commit augments and leverages that macro to remove class_finalize function definition boilerplate.
* daemon: Remove unused UPOWER_* variablesBastien Nocera2012-06-151-2/+0
|
* daemon: Fix g_clear_object() usageDaiki Ueno2012-05-231-5/+2
| | | | https://bugzilla.gnome.org/show_bug.cgi?id=673555
* daemon: don't unref member objects twice on disposeDaiki Ueno2012-05-171-4/+4
| | | | https://bugzilla.gnome.org/show_bug.cgi?id=673555
* daemon: Drop dbus activation capabilityRay Strode2012-04-053-67/+8
| | | | | | | | | | | | | | | | | | | | | | | | | Years ago, GDM activated gnome-settings-daemon via dbus as part of its hand-rolled greeter session. These days it uses gnome-session instead of a hand-rolled session, and so gnome-settings-daemon no longer needs to be activatable. Furthermore, when gnome-settings-daemon is activated, the only way it functions is if it's subsequently given an "Awake" call. This limitation was added because it was getting unintentionally activated in none-GNOME sessions (when those sessions were hosting gnome apps, like rhythmbox), causing all sorts of havoc. This doesn't-function-until-Awake-called limitation means that if settings daemon ever crashes and gets unintentionally dbus activated before gnome-session has a chance to respawn it, then the whole desktop breaks (since whatever accidentally activates it won't call Awake). This actually happens in practice because gnome-shell's power indicator tries to track the power plugin's object properties. This commit removes the ability for gnome-settings-daemon to be dbus activated. It only causes problems, and no longer solves anyway. https://bugzilla.gnome.org/show_bug.cgi?id=673586
* Simplify previous commitMartin Pitt2012-03-131-2/+2
|
* daemon: Support explicitly setting G_MESSAGES_DEBUGMartin Pitt2012-03-121-1/+1
| | | | | | | | Do not force G_MESSAGES_DEBUG to "all" if it is already set. This makes it much easier to limit debug output to a particular plugin by starting g-s-d with e. g. G_MESSAGES_DEBUG="mouse-plugin". https://bugzilla.gnome.org/show_bug.cgi?id=671928
* main: Fix library linkage on MageiaJani Välimaa2012-03-091-0/+1
| | | | | | libgsd.so refused to link as some glib symbols didn't get defined. https://bugzilla.gnome.org/show_bug.cgi?id=670651
* build: Remove empty variablesBastien Nocera2012-03-091-13/+0
|
* main: Drop soname from the private libraryBastien Nocera2012-02-231-2/+4
|
* main: Fix problems when libexecdir == libdirBastien Nocera2012-02-231-1/+3
| | | | | | As done in Debian and OpenSUSE https://bugzilla.gnome.org/show_bug.cgi?id=670695
* build: Fix build with --enable-systemdBastien Nocera2012-02-211-1/+1
| | | | | | Link the shared library that uses it with systemd's libraries https://bugzilla.gnome.org/show_bug.cgi?id=670525
* Unify libgsd-profile and libgsd-session into installed shared libraryColin Walters2012-02-181-28/+8
| | | | | | | | | | | | | | Currently plugins can call back into the main binary; if the program is linked with -Bsymbolic then what happens is the plugin process will contain duplicate functions from libgsd-session.la. Ubuntu has patched their toolchain to link with -Bsymbolic-functions by default which causes a failure here. This is also just cleaner and better; there's only one copy of the code on disk. https://bugzilla.gnome.org/show_bug.cgi?id=670244
* daemon: Remove deprecated g_thread_init()Bastien Nocera2012-02-081-6/+0
|
* Use GnomeSettingsSession in the automounterMatthias Clasen2012-01-231-2/+21
| | | | | | This avoids duplicating the session monitor code in two places. https://bugzilla.gnome.org/show_bug.cgi?id=666787
* Add a systemd implementation of GnomeSettingsSessionMatthias Clasen2012-01-232-23/+152
| | | | | | | | | This implementation uses the logind apis for tracking active sessions instead of ConsoleKit. It is only compiled if --enable-systemd is passed to configure, otherwise the ConsoleKit implementation is used. https://bugzilla.gnome.org/show_bug.cgi?id=666787
* Merge branch 'docking-stations' - bgo#663126Federico Mena Quintero2012-01-103-19/+2
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This lets docking stations work better in the light of plugging external monitors. It used to be that a docked laptop, with a closed lid, would still keep the LCD active, thus making it pretty useless with an external monitor. Now, we disable the LCD if the laptop's lid is closed, as expected. Additionally, to decide when to suspend the laptop, we do so if there are no plugged or usable RANDR outputs. For example, a docked laptop with a closed lid and *no* external monitor has no usable outputs, so it will get suspended. This requires gnome-desktop 3.3.4 or later, as the semantics of gnome_rr_screen_new() are slightly different there - now it returns a singleton instead of a new object every time.
| * Remove gnome_settings_session_get_upower_client()Federico Mena Quintero2012-01-102-19/+0
| | | | | | | | | | | | | | It was done to return a singleton from up_client_new(), but in fact up_client_new() already returns a singleton. Signed-off-by: Federico Mena Quintero <federico@gnome.org>
View Lorry Controller Status